NMR relaxation of neritic carbonates: An integrated petrophysical and petrographical approach
Résumé
A set of carbonate outcrop samples, covering a wide range of the sedimentary textures and depositional environments existing on carbonate systems, was studied through an integrated petrographical and petrophysical approach. With the aim of improving the understanding of the NMR (Nuclear Magnetic Resonance) signal of carbonates, this work is: 1) providing an atlas for various carbonate reservoir rock-types, 2) providing a workflow for integrating geological and petrophysical data and, 3) documenting common shortfalls in NMR/MICP analyses in carbonates. The petrographical investigation includes thin section and SEM (Secondary Electron Microscope) observations, whereas petrophysical investigation includes porosity (Φ), permeability (K), NMR, MICP (Mercury Injection Capillary Pressure), and specific surface area (BET) measurements. On the basis of NMR and MICP data, 4 groups of samples were identified: (1) microporous samples, (2) micromesoporous samples, (3) wide multimodal samples, and (4) atypical samples. The microporous samples allow us to define a maximum NMR threshold for microporosity at a T 2 of 200 ms. NMR and MICP response of the investigated carbonates are often comparable in terms of modal distribution (microporous, micro-mesoporous and wide multimodal samples). In particular, micritization, a well known but underestimated early diagenetic process, tends to homogenize the NMR signal of primarily different sedimentary facies. A grainstone with heavily micritized grains can display well sorted unimodal NMR and MICP signatures very similar, even identical, to a mudstone-wackestone. Their signatures are comparable to that of a simple sphere packing model. On the contrary, several samples (labeled atypical samples) show a discrepancy between NMR and MICP response. This discrepancy is explained by the fact that MICP can be affected by the physical connectivity of the pore network, in case of disseminated and isolated molds in a micrite matrix for instance. Similarly, NMR can differentiate pore classes but with less resolution. It does not rely on connectivity but can be affected by diffusional pore coupling, i.e. the diffusion of water molecules carrying the magnetization between micropores and macropores. The pore-coupling phenomenon, through its impact on the T 2 distribution, may disturb the permeability calculations from NMR data. For core plug characterization, NMR appears to be a complementary tool to MICP and should be used to complete rock-typing analysis in carbonates. For reservoir rock-typing, the obvious advantage of developing a NMR based approach is the use of NMR logging data providing continuous records of pore size distributions.